.386 METALLURGY OF IRON AND STEEL.
steel gave a value for carbon of 1000 Ib. per unit of 0.01 per cent,, and as the basic steel gives 810 Ib. when calculated by the acid formula and 770 Ib. by its own formula, it would seem certain that a unit of carbon has much less effect upon basic than upon acid steel.
THE APPLICATION" OF THE FORMULA.
Table XVII-Y shows the result of comparing the actual strength of the steels under consideration with the strength as calculated from the formulae just given. For this purpose the heats were grouped according to carbon and then subdivided according to manganese. No heats were put together that varied more than 0.05 per cent, in carbon, or more than 0.1 per cent, in manganese. For instance, a group might include a heat containing 0.1 per cent, of carbon and 0.3 per cent, of manganese, and another heat containing 0.149 per cent, of carbon and 0.399 per cent, of manganese, but any heat of higher or lower carbon, or of higher or lower manganese than these extremes, would fall into another group. Inasmuch as the phosphorus did not vary through wide limits in any of the steels, each group may be looked upon as composed of heats that are • practically alike in chemical composition, and which may properly be averaged to eliminate accidental errors.
In some of the subdivisions the number of heats is so small that these errors cloud the result. Especially in the steels of higher carbon it is desirable to have a large number of heats in the average, as it is difficult to get uniform results on a testing-machine under usual working conditions when the bar has a strength of over 90,000 Ib. per sq. in., and unfortunately it is in these high steels, and particularly in the groups with an unusual content of manganese, that only a small number of heats were on record. There are, accordingly, several instances where these sm'all groups show a considerable difference between the actual and the calculated strength, but there seems to be no rule as to the difference, as other groups, either large or small, of the same class of steels give satisfactory results.
It is, of course, a matter of opinion as to what constitutes a fair agreement between the actual and the calculated strengths, but in the following comparison, it will be assumed that the results of the formulae should be within 1500 Ib. of the records of the testing-